Grok 3 was a very good model, and Grok 4 might be amazing but having a very good model is not enough – there are a lot of really good models out there. You actually want to trust the model you are building on.
LLMS
-
Half Usage Split Between Direct Cline and OpenRouter Integration
By
–
I know 🙁 ~half my usage is here, the other half is directly in Cline via @openrouter
-
Is Llama Dead? Polymarket Bet on Llama 5 by 2026
By
–
so like… Llama is dead dead right? can someone start a polymarket on whether there will be a llama 5 by 2026
-
SmolLM3: Powerful built-in tool-calling capabilities
By
–
Reminder: SmolLM3 comes with built-in tool-calling, and it works really well! pic.twitter.com/uba78UfPtg
— m_ric (@AymericRoucher) 9 juillet 2025Reminder: SmolLM3 comes with built-in tool-calling, and it works really well!
-
Which AI Model and Advanced Features Did You Use?
By
–
What model did you use? Did you turn on Thinking? Did you use Plan/Act?
-
Causal Masking: The Unsung Hero Behind Transformer Success
By
–
an unsung hero of the Transformer's success is not the attention mechanism per se, but *causal masking* it let us simplify two for loops to just one: > RNN
for t in [1..S]: for i in [1..L]: x[t], h = layers[i](x[t], h) > transformer for i in [1..L]: x = layers[i](x) -
FlashAttention less useful with MLP GEMM latency dominance
By
–
Maybe FlashAttention is not that useful when you have MLP GEMMs that eat so much latency? Interesting graph in the latest blog post from @gpus_go_brrr!
-
Gemini 3.0 Pro reference discovered in Gemini-CLI commit
By
–
BREAKING : Gemini 3 reference has been spotted in the Gemini-CLI commit! gemini-beta-3.0-pro
-
OpenAI to Release New Open-Weight Model Matching o3-mini Performance
By
–
BREAKING : According to The Verge, OpenAI will release its open weight model next week and it will be performing at the o3-mini level.
-
ASC+ Becomes New AI Standard Default
By
–
Everything is ASC+ now unless you specifically tell it not to. I'm almost always referring to ASC+ as this is the standard now.